Todd Vierling <tv@pobox.com> writes:
> On Wed, 9 Sep 1998, Ryan Ordway wrote:
> : > Um, that shouldn't make sshd, for instance, core-dump.  If that was
> : > happening, it dounds like (1) the device lookup code for /var/db/kvm.db is
> : > core-dumping or (2) kvm_mkdb was making a bogus file.  In either case, I
> : > think something is broken there, and I never really liked the Berkeley-DB
> : > format kvm.db.  :>
> : 
> : 	Yeah, kvm_mkdb dumped core first, which I am guessing is why the
> : others dumped core. If someone would like, I can put it up for anon ftp.
> 
> Probably not a big matter, but the kvm_db stuff really needs to be rethought
> (that was my original point :).

huh?  what are you talking about?

approximately _none_ of those programs (other than kvm_mkdb;
specifically i'm thinking: sshd, getty, sh, etc.) should be using the
kvm database.

instead, you'll note that they are all likely to use db bits, which
have recently been updated to use pread, etc.

I'd be interested to know:

* what versions of kernels (and with what names) worked with what
versions of libraries

I.e. an explicit list.

To be quite honest, the lossage sounds like it could be from using a
new libc with an old kernel, rather than anything else.  (that
shouldn't be generating seg-faults, though, I don't think.)


cgd
-- 
Chris Demetriou - cgd@netbsd.org - http://www.netbsd.org/People/Pages/cgd.html
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.
Plug: Get your official NetBSD-1.3.2 CDROM set today! http://www.netbsd.com/